cancel
Showing results for 
Search instead for 
Did you mean: 

STM32G0 Custom Board NRST floating

BKN636
Associate II

Hello everyone

ive been using STM32 microcontroller on custom boards for some while now, but particularly with the STM32G071G8 some boards are failing. After a few minutes, NRST seems to flicker and drop from 3V3 to 2,2, 1,7 and finally steady to around 400mV. µC cant be programmed due to its being held in constant reset state.

For hardware design ive been using this guide from STM.

Things ive tried:

  • thoroughly cleaning the board with ethanol and PCB cleaner to make sure no residual solder and flux causing shorts on the pins
  • All power pins are bypassed with 100nF and 4,7µF in parallel, even tried an additional 10µF tantal cap
  • Using an external 4k7 Pullup to 3V3 didnt help either.
  • Reading the processor reset flags over the COM port states an external reset.
  • letting the boards unpowered for a few days
  • "reviving" the controller with a few seconds of hot air
  • board variants with the standart footprint recommended by STM for the UFQFPN28 package and and custom footprint with elongated pads for easier hand soldering - no improvements

Here are some schematics and board screenshot of my custom boards:

0693W000005A03FQAS.png0693W000005A03AQAS.png0693W000005A0GnQAK.png0693W000005A035QAC.png

5 REPLIES 5
TDK
Guru

That chip has an exposed pad, and you have an exposed NRST via under the pad. It may work, but could be that pad is touching the NRST via.

You can still program the chip if you select "Connect under reset".

Measure connectivity between NRST and GND.

If you feel a post has answered your question, please click "Accept as Solution".
BKN636
Associate II

As far as i can tell that µC doesnt have an exposed pad, so routing underneath with NRST shouldnt be an issue.

Resistance of a freshly soldered µC between NRST and GND is somewhere in the MOhms Region. If the pin is floating its near 2k-3k, measured while the board is unpowered ofc

You're right. It seems like the UFQFPN28 doesn't have an exposed pad, but the larger UFQFPN packages do. I was going off of Mouser's picture.
Does the 3V3 rail remain steady while NRST falls? Can you get a scope on NRST to see if it's actually dropping or just getting reset repeatedly?
If you feel a post has answered your question, please click "Accept as Solution".
BKN636
Associate II

3V3 Supply looks steady.

NRST looks a bit like this:

0693W000005A4dHQAS.png

Uwe Bonnes
Principal III

The time scale is missing. Do you have several boards to test and compare.